Genetic variation in endocannabinoid signaling: Anxiety, depression, and threat- and reward-related brain functioning during the transition into adolescence

Abstract

Background:

The endocannabinoid system modulates neural activity throughout the lifespan. In adults, neuroimaging studies link a common genetic variant in fatty acid amide hydrolase (FAAH C385A)—an enzyme that regulates endocannabinoid signaling—to reduced risk of anxiety and depression, and altered threat- and reward-related neural activity. However, limited research has investigated these associations during the transition into adolescence, a period of substantial neurodevelopment and increased psychopathology risk.

Methods:

This study included FAAH genotype and longitudinal neuroimaging and neurobehavioral data from 4,811 youth (46% female; 9–11 years at Baseline, 11–13 years at Year 2) from the Adolescent Brain Cognitive Development^SM Study. Linear mixed models examined the effects of FAAH and the FAAH x time interaction on anxiety and depressive symptoms, amygdala reactivity to threatening faces, and nucleus accumbens (NAcc) response to happy faces during the emotional n-back task.

Results:

A significant main effect of FAAH on depressive symptoms was observed, such that depressive symptoms were lower across both timepoints in those with the AA genotype compared to both AC and CC genotypes (p’s<0.05). There were no significant FAAH x time interactions for anxiety, depression, or neural responses (p’s>0.05). Additionally, there were no main effects of FAAH on anxiety or neural responses (p’s>0.05).

Conclusions:

Our findings add to emerging evidence linking the FAAH C385A variant to lower risk of psychopathology, and extend these findings to a developmental sample. In particular, we found lower depressive symptoms in FAAH AA genotypes compared to AC and CC genotypes. Future research is needed to characterize the role of the FAAH variant and the eCB system more broadly in neurodevelopment and psychiatric risk.

1. Introduction

The endocannabinoid (eCB) neuromodulatory system consists of cannabinoid type 1 receptors, which are densely located in limbic brain regions like the amygdala and nucleus accumbens (NAcc) [1,2], and endogenous ligands (eCBs) such as N-arachidonoylethanolamine (anandamide, AEA) and 2-arachidonoylglycerol (2-AG). AEA and 2-AG are lipid mediators produced on-demand in response to stress and act in a retrograde fashion to reduce pre-synaptic activity [3]. Enzymes like fatty acid amide hydrolase (FAAH) degrade eCBs, regulating their concentrations in the synapse. The eCB system is often studied using a common functional polymorphism in the FAAH gene: FAAH C385A (rs324420). This polymorphism substitutes proline with threonine at amino acid position 129. Studies consistently show that individuals carrying the FAAH A-allele have higher concentrations of circulating AEA than those with the CC genotype [4,5].

Recent studies link FAAH C385A to altered risk of stress-related neuropsychiatric disorders, including anxiety, depression, and posttraumatic stress disorder. Indeed, both preclinical and clinical studies associate the FAAH A-allele with lower anxiety and better fear extinction—a process relevant to the pathophysiology and treatment of anxiety [6]—as compared to the CC genotype [7–9]. Conversely, studies have also reported higher depressive symptoms in those with the FAAH A-allele compared to the CC genotype [4,10]. Together, these studies suggest that FAAH C385A is associated with altered risk of stress-related disorders; however, the direction of results remains unclear and may vary across disorders or populations.

Few studies have examined the impact of the FAAH C385A variant on threat and reward processing neural circuitry, which are relevant to the pathophysiology and treatment of stress-related neuropsychiatric disorders [11]. Central to reward circuitry is the NAcc, which is located within the ventral striatum and plays a significant role in goal-directed behaviors, behavioral sensitization, and alterations in affective states [12]. For example, lower reward-related NAcc activity is implicated in vulnerability to depression [13]. Correspondingly, the amygdala is a key brain region involved in threat processing and the acquisition and expression of conditioned fear. Prior studies demonstrate higher threat-related amygdala reactivity in individuals with anxiety and depression as compared to healthy controls [14,15].

A prior study by Hariri and colleagues (2009) used functional magnetic resonance imaging (fMRI) to investigate the impact of FAAH C385A on threat- and reward-related neural responses in a community sample of 82 adults without psychiatric disorders [16]. The authors found higher reward-related activity in the ventral striatum, but lower threat-related reactivity in the amygdala, among A-allele carriers as compared to their CC genotype counterparts. There was also a positive association between amygdala threat reactivity and trait anxiety in adults with the CC genotype (but not A-alleles). These results are consistent with studies examining the relationship between FAAH C385A or FAAH levels in brain (using the FAAH [¹¹C]CURB radiotracer) and fear extinction or threat processing. For example, one positron emission tomography study in healthy adult volunteers reported an association between higher binding of the [¹¹C]CURB tracer (suggesting higher levels of FAAH) in the amygdala and higher amygdala fMRI response to threatening social cues [17]. This finding aligns with preclinical and neuroimaging studies in humans that suggest a role for FAAH in modulating stress and anxiety. Additionally, prior studies have linked the FAAH A-allele to quicker amygdala habituation to threat and lesser amygdala activation during a test of extinction recall [18,19]. For reward, Hariri et al. found a negative correlation between ventral striatal reward-related activity and delay discounting, a measure of impulsivity, in CC genotypes whereas A-alleles showed a positive association [16]. Together, these findings suggest divergent effects of the FAAH variant on threat- vs. reward-related functioning.

In a second study, also in a community sample of adults without psychiatric disorders (n = 112), Conzelmann et al. [20] examined the impact of FAAH C385A on affect-modulated startle reflex during the presentation of pleasant, neutral, and unpleasant images from the International Affective Picture System [20]. In this study, A-alleles demonstrated increased startle potentiation—indicating an increased reactivity—towards unpleasant pictures, which the authors suggest reflects exaggerated threat-related amygdala activity. A-alleles in the study also demonstrated reduced startle inhibition—reflecting reduced reactivity—towards pleasant (rewarding) pictures, suggesting reduced reward-related NAcc activity. Although Conzelmann et al. did not use neuroimaging, their startle results contrast with the earlier study by Hariri et al. In sum, although the two existing studies in adults are conflicting and relatively limited in sample size (n’s < 120), they are the first to link the FAAH C385A variant to altered threat- and reward-related brain functioning and potential risk of stress-related disorders.

Stress-related disorders frequently begin during the transition from childhood to adolescence (~11 years) [21]. However, few studies have investigated the impact of the FAAH C385A variant on psychiatric symptoms or brain functioning during this important developmental window, when threat- and reward-related neural circuitry undergo dynamic restructuring [22,23]. Only three studies have examined the effects of FAAH C385A on psychiatric symptoms or neuroimaging outcomes in children or adolescents [24–26]. In a cross-sectional study, Gee et al. reported higher fractional anisotropy—a nonspecific measure of white matter integrity—within frontolimbic circuitry and lower anxiety symptoms in individuals with the A-allele (compared to the CC genotype) during adolescence (12–18 years), but not preadolescence (6–11 years) [25]. In another study, higher frontolimbic resting-state connectivity was found in preadolescent children (ages 9–11) with A-alleles compared to CC genotypes, but no main effects of the FAAH C385A polymorphism on anxiety were observed [24]. In a third study, lower fractional anisotropy was observed in the left fornix and left parahippocampal cingulum in preadolescent children with A-alleles (vs. CC genotypes), and FAAH moderated the association between fractional anisotropy of the left cingulum bundle and anxiety symptoms [26]. These studies suggest that main effects of the FAAH C385A variant on anxiety and frontolimbic circuitry may emerge during adolescence, and are not yet present in preadolescence. However, no studies have examined associations among FAAH, threat- and reward-related neural responses, and anxiety or depressive symptoms in youth.

Here, we used longitudinal data from the Adolescent Brain Cognitive Development^SM Study (ABCD Study^®) to characterize the effects of the FAAH C385A variant on anxiety and depressive symptoms and threat- and reward-related neural responses during the transition into adolescence. The overall aim of this study was to replicate and extend the results of Hariri et al. [16] to the transition into adolescence, with a larger sample size, and by incorporating a longitudinal design. We used a region of interest (ROI) approach to focus a priori on amygdala response to threat and NAcc response to reward. Based on prior studies [21,27], we predicted an overall increase in anxiety and depressive symptoms (i.e., main effect of time) from preadolescence (~9–11 yrs) to adolescence (~11–13 yrs). Although prior studies in adults are contradictory [16,20], we also predicted that the FAAH C385A variant would be associated with lower anxiety and depression during adolescence but not preadolescence (i.e., FAAH x time interaction). These predictions were based on the existing studies in developmental samples linking the FAAH C385A variant to lower anxiety during adolescence (but not preadolescence) [4,24,25]. Additionally, we predicted lower amygdala threat-related activity but higher NAcc reward-related activity in A-alleles (but not CC genotypes) during adolescence but not during preadolescence. This prediction is based on the only fMRI study to our knowledge to examine the relationship between the FAAH variant and threat- and reward-related processing [16], and that neurobiological effects of the FAAH C385A variant may emerge during adolescence [28,29]. Symptom-brain correlations were also explored (e.g., FAAH x brain response in predicting symptomology), given Hariri et al.’s observed negative correlation between amygdala reactivity and trait anxiety, and positive correlation between ventral striatal reactivity and delay discounting (a marker of behavioral impulsivity and/or reward sensitivity) in FAAH A-alleles only [16].

2. Methods and Materials

2.1. Participants

The present study included genotype data as well as longitudinal neuroimaging and neurobehavioral data from the ABCD Study^®. Data were accessed through the NIMH Data Archive (NDA) Release 4.0 (dx.doi.org/10.15154/1523041). Written informed consent and verbal assent was obtained from parents/guardians (“caregivers”) and children, respectively. The study was approved by the central and local institutional review boards. Full details of ethics, oversight, inclusion/exclusion criteria, and study procedures for the ABCD Study^® can be found elsewhere [30–32].

Here, we used data from Baseline (9–11 years old) and Year 2 follow-up (11–13 years old) visits. Our final sample consisted of 4,811 youth with available (1) genotype, (2) Baseline and Year 2 anxiety and depressive symptoms, and (3) Baseline and Year 2 fMRI data. We excluded participants whose genetic data did not pass quality inspection, as reported in the Known Issues with Data Release 4.0. Detailed demographic data, separated by each genotype, can be found in Table 1.

Table 1.

Participant demographics.

	Overall (n = 4,811)	FAAH AA (n = 304)	FAAH AC (n = 1,718)	FAAH CC (n = 2,789)	p-value
Age in years, M (SD)
Baseline	9.93 (0.62)	9.94 (0.61)	9.91 (0.63)	9.94 (0.62)	p=0.462
Year 2	11.92 (0.64)	11.93 (0.63)	11.90 (0.65)	11.93 (0.64)	p=0.337
Biological Sex, n (%)					p=0.8
Female	2,263 (47.04%)	139 (45.13%)	802 (46.68%)	1,322 (47.4%)
Male	2,548 (52.96%)	165 (53.57%)	916 (53.32%)	1,467 (52.6%)
Puberty Score, M (SD)
Females:
Baseline	1.74 (0.52)	1.77 (0.56)	1.77 (0.53)	1.72 (0.50)	p=0.103
Year 2	2.47 (0.66)	2.46 (0.69)	2.51 (0.66)	2.44 (0.65)	p=0.046
Males:
Baseline	1.42 (0.37)	1.49 (0.41)	1.46(0.40)	1.40 (0.35)	p<0.001
Year 2	1.78 (0.55)	1.86 (0.54)	1.80 (0.56)	1.77 (0.56)	p=0.112
Race, n (%)					p<0.001
White	2859 (59.43 %)	135 (44.41%)	926 (53.90%)	1798 (64.47%)
Non-White	1952 (40.97%)	169 (55.59%)	792 (46.10%)	991 (35.53%)
Combined Household Income, n (%)					p<0.001
<$50,000	1,140 (25.52%)	83 (30.18%)	476 (29.96%)	581 (22.34%)
$50,000– 99,999	1,381 (30.92%)	91 (33.09%)	477 (30.02%)	813 (31.26%)
$100,000+	1,946 (43.56%)	103 (37.45%)	636 (40.03%)	1,207 (46.4%)

Note that all participants did not report all demographics, and some data are missing. P-values indicate differences between FAAH genotypes. Puberty scores were measured by calculating the sum score of the Puberty Development Scale [74].

2.2. Genotyping

Smokescreen_TM Genotyping array was used to define genotypes using salvia samples collected from participants [32]. Further details on DNA sample collection and processing can be found elsewhere [31]. Genotyping data were used to examine the FAAH C385A polymorphism. Participants were characterized with three mutually exclusive groups: AA, AC, and CC genotypes.

2.3. Anxiety and depression symptoms

Anxiety and depressive symptoms were measured using the Child Behavior Checklist (CBCL) [33]. The CBCL is a widely used caregiver-reported measure of dimensional psychopathology and adaptive functioning. We used the DSM-5 oriented age- and sex-normed (T-scores) for anxiety and depressive disorder scales [30].

2.4. Neuroimaging

Task-based fMRI data were collected during the emotional n-back test, which is assessment of working memory function that measures the participant’s ability to direct attention away from the emotional faces to focus on a cognitive task [30]. Rather than isolating neural response to working memory, we focused on tabulated fMRI contrast (i.e., activation) values from the NDA for amygdala response to negative (i.e., threatening faces) versus neutral faces and NAcc activation to positive (i.e., happy faces) versus neutral faces [32]. This was done to isolate neural response to threat-related and reward-related processing [34]. Beta weights for n-back task fMRI response in subcortical ROIs were averaged across all runs; here, we focused on fMRI response in anatomical amygdala and NAcc ROIs [32,35]. Considering the results of Hariri et al. [16] and known hemispheric lateralization of emotion-related processing [36,37], analyses were conducted in the left and right hemisphere, separately for a total of four ROIs: left amygdala, right amygdala, left NAcc, right NAcc. Further details on imaging data acquisition and preprocessing are provided in other sources [32].

2.5. Statistical analyses and covariates

Linear mixed effects models using three genotype-specific comparisons (i.e., AA vs. CC, AC vs. CC, and AA vs. CC) were used to test for main effects of FAAH and the FAAH x time interaction on (1) symptoms and (2) neural activation. For (1), separate models were performed for anxiety and depressive symptoms. For (2), separate models were performed for right and left amygdala reactivity to threat and right and left NAcc reactivity to reward. Finally, we explored (3) symptom-brain associations. For (3), linear mixed effect models were used to test for FAAH x ROI interactions in predicting anxiety and depressive symptoms. See Figure 1 for a summary of analyses. Statistical analyses were performed in R Studio (RStudio-2023.06.1–524) using lme4 to calculate linear mixed effects models and were adjusted for the following covariates and random effects: number of adverse childhood experiences (ACEs), annual family income, age at Baseline (analysis 2 only), sex (analysis 2 only), and race. Additionally, family identity and site identity were included as random effects, with family nested within site. Covariates were determined using a directed acyclic graph (DAG) to identify potential confounding pathways (see Figure 1) [38,39], and given their associations with outcome measures in this study (see Supplemental Material). ACEs were measured using the total number of traumatic events endorsed on the PTSD Criterion A Traumatic Events Screener and the Kiddie Schedule for Affective Disorders and Schizophrenia (K-SADS) for DSM-5 [40]. SES was measured using caregiver-reported combined household income (see Table 1). P-values and 95% confidence intervals (CI) are reported.

Figure 1.

Directed acyclic graph (DAG) outlining the predicted associations among variables of interest in this study, including analyses of interest (1–3) and potentially confounding pathways. Confounding variables are indicated in grey; the independent variable (FAAH C385A variant) is in green; dependent variables (brain and symptoms) are given in blue. We examined the impact of the FAAH C385A variant on (1) anxiety and depressive symptoms and (2) threat- and reward-related brain functioning in the amygdala and NAcc, respectively. We also explored (3) the interaction between FAAH and threat-and reward-related brain functioning in predicting anxiety or depressive symptoms. FAAH, fatty acid amide hydrolase; NAcc, nucleus accumbens.

3. Results

3.1. Effects of the FAAH C385A variant on anxiety and depressive symptoms

There was a significant main effect of FAAH on depressive symptoms (Figure 2A) such that AA genotypes had lower depressive symptoms across both time points compared to both AC (β=−0.83 95% CI: −1.55 to −0.10; p=0.025) and CC (β=−0.75, 95% CI: 0.03 to 1.47, p=0.042) genotypes. There were no group differences in depressive symptoms between AC and CC genotypes (β=−0.052 95% CI: −0.41 to 0.31; p=0.78). There was no significant FAAH x time interaction on depressive symptoms (p’s>0.2), and no main effects of FAAH, or FAAH x time interactions (p’s>0.05) on anxiety symptoms (Figure 2B).

Figure 2. Violin plot showing depressive (A) and anxiety (B) symptoms over time, by FAAH C385A genotype.

There was a significant main effect of FAAH such that FAAH AA alleles had lower depression scores on average compared to AC and CC genotypes across time points. Error bars represent standard error. *p<0.05. Means shown are not adjusted for covariates. FAAH, fatty acid amide hydrolase.

3.2. Effects of the FAAH C385A variant on threat- and reward-related neural activity

There were no significant main effects of FAAH on left or right threat-related amygdala or left and right reward-related NAcc activity (p’s>0.05). There were no significant FAAH x time interactions for any ROI (p’s>0.2; Figure 3A–D).

Figure 3. Violin plot showing reward-related NAcc and threat-related amygdala reactivity from Baseline to Year 2, by FAAH C385A genotype.

(A) Left NAcc reward response. (B) Right NAcc reward response. (C) Left amygdala reactivity to threat. (D) Right amygdala reactivity to threat. There were no significant main effects of FAAH, time, or FAAH x time interactions. Error bars represent standard error. Means shown are not adjusted for covariates. FAAH, fatty acid amide hydrolase; NAcc, nucleus accumbens.

3.3. Exploratory: Symptom-brain associations

There were no significant neural activity x FAAH genotype interactions on predicting anxiety or depressive symptoms (p’s>0.05).

4. Discussion

To our knowledge, this is the first longitudinal investigation of the FAAH C385A variant and its associations with anxiety and depressive symptoms, as well as threat- and reward-related neural reactivity, in a developmental sample. We used longitudinal data from the ABCD Study^® from Baseline to Year 2, which captures the transition from preadolescence into adolescence. Two main results emerged: First, youth with the FAAH AA genotype showed lower depressive symptoms compared to both AC and CC genotypes across both time points (i.e., Baseline and Year 2). Second, anxiety symptoms, NAcc reward-related activity, and amygdala threat-related reactivity were not associated with FAAH (i.e., no main effects or interactions with time). Exploratory analyses did not demonstrate a link between FAAH genotype x brain response in predicting anxiety or depressive symptoms. Overall, our findings suggest that the potential protective effects of the FAAH C385A variant on depression are present prior to the transition into adolescence. These results extend prior work in adults and in cross-sectional developmental samples, but do not support effects of the FAAH C385A variant on amygdala threat-related reactivity or NAcc reward-related activity in this developmental sample.

Overall, we found that FAAH AA genotypes had lower depressive symptoms compared to AC and CC genotypes. The transition into adolescence is characterized by pronounced biological and social changes, including hormonal fluctuations, neurogenesis, dendritic growth, and substance use [41,42]. Depressive symptomology and reported rates tend to be low during preadolescence [43] and rise substantially throughout adolescence [44]. The present study is the first to our knowledge to report an association between FAAH and depressive symptoms in a developmental sample. However, the current literature in adults shows conflicting results for the relationship between FAAH and depressive symptoms. Contrary to our findings, previous studies in adults report higher depressive symptoms in those with the FAAH AA genotype [4,10]. In addition, one study did not find a significant association between FAAH and depressive symptoms [45]. However, consistent with the present results, a review of studies in rodent models supports a link between lower FAAH gene expression and protein levels (as expected in AA genotypes), and lower depressive-like behaviors [46]. In addition, the prior study by Gee and colleagues reported lower anxiety symptoms—which often co-occur or precede the onset of depression [47]—in adolescents carrying the FAAH A-allele compared to CC genotypes [25]. Future neuroimaging studies are needed to replicate and extend these results in developmental samples.

We did not observe any effects of the FAAH C385A on reward-related NAcc reactivity overall, or across time. In typically-developing populations, prior studies report a normative increase in NAcc activity during the transition from preadolescence to adolescence [48–51], with a peak in mid-adolescence [52]. Here, we did not find an overall change in NAcc activity to happy faces during the emotional n-back task from preadolescence to adolescence. Our null findings for the effects of the FAAH C385A genotype on NAcc activity differ from the prior fMRI study by Hariri et al. [16] in adults. Hariri and colleagues found lower reactivity of the ventral striatum, which contains the NAcc, in CC genotypes vs. A-alleles during a gambling task with positive and negative feedback and the possibility to win money [16]. These conflicting results may reflect the protracted development of the NAcc and broader reward circuitry and age-related changes in NAcc response to monetary and social rewards from preadolescence to adolescence [53], or task-related differences (e.g., happy faces as a social reward vs. gambling task). However, a later study in adults by Conzelmann et al. [20] reported an opposing pattern of increased startle inhibition in CC genotypes relative to A-alleles, indicating increased emotional reactivity towards pleasant pictures. Taken together, these findings suggest that the effects of the FAAH C385A variant on emotional processing are complex and may differ across development.

Surprisingly, we found no effects of time or FAAH on anxiety symptoms or threat-related amygdala reactivity at Baseline or at Year 2. Anxiety symptoms tends to increase in prevalence and severity in early adolescence [54,55] and persist into and throughout adulthood if left untreated [56]. Prior cross-sectional studies in both adults and youth link the FAAH CC genotype to elevated anxiety symptoms and higher threat-related amygdala reactivity as compared to the A-allele [16,25]. However, prior studies in youth (~9–11 yrs) using Baseline ABCD Study^® data also did not report significant main effects of FAAH on anxiety or frontolimbic connectivity [24,26]. These null effects suggest that effects of FAAH on anxiety emerge later, during mid-to-late adolescence, or during adulthood. Alternatively, effects of FAAH on anxiety or amygdala reactivity may be more complex and involve interactions with ACEs, or via modulation of symptom-brain associations [cf. 24,26]. FAAH may also have effects on other outcomes that were not examined here, such as pain or inflammatory response [57–59]. Future longitudinal waves of the ABCD Study^® will be poised to address these effects.

4.1. Strengths and Limitations

The current study contains many strengths, including the large, multi-site neuroimaging study, longitudinal approach, and attempted replication and extension of a prior study in adults. However, this study is not without limitations. It should be noted that test-retest reliability of task-based fMRI studies can be poor [60–62]. One study attributed poor emotion processing reliability to emotional valence not explicitly queried as part of the n-back task used in the ABCD Study^® [63]. However, the authors of that study also point out that the large sample size afforded by the ABCD Study^® provides sufficient statistical power to detect significant correlations even with low reliability/stability traits [63]. Additionally, although the emotional n-back task is a working memory task, it has been shown to engage emotion processing regions in both fronto-amygdala and fronto-striatal circuitry [64,65]. Nonetheless, other tasks that directly measure threat or reward may be beneficial for examining the effects of the FAAH variant on neural response. Additionally, we observed a relatively small effect size for the main effects of FAAH on depressive symptoms (β=−0.8). While large sample sizes can provide increased statistical power to detect smaller effect sizes to significance, these samples can pose additional challenges for assessing real-world significance [24]. However, as pointed out by others, even small effect sizes when extrapolated to larger populations can have large publication health implications [66]. For example, if CC genotypes have increased depressive scores by ~0.8 at year 2 on average, approximately 40 out of every 1,000 children would move from at risk to within the clinical range for depression within 4–6 years based on CBCL criteria. While we included the number of ACEs as a covariate in our statistical models, it was beyond the scope of the current paper to explore interactive effects of FAAH genotype with ACEs or other environmental factors (e.g., neighborhood quality, substance use) [67,68]. Future studies should explore these effects, as well as other genetic variants such as cannabinoid type 1 receptor genotype (CNR1), given these and others may have additional or interactive effects on brain and behavioral outcomes [69]. In addition, the reliance on parent-reported measures of youth anxiety and depression is a limitation [70,71]. Lastly, the assessment of FAAH is an indirect measure of eCB system functioning. Although the FAAH A-allele is typically associated with higher circulating eCB concentrations as compared to the CC genotype, there can be substantial within-genotype variability [72]. Given that eCB concentrations have been implicated in risk of anxiety and depression [73], future research should incorporate measures of circulating eCBs (e.g., plasma).

5. Conclusion

Our findings suggest that the protective effects of the FAAH C385A variant are present prior to and during transition into adolescence. The eCB system is crucial for the development of functional brain networks [28] and is linked to risk of stress-related disorders, which often begin during adolescence [27]. Additional research is necessary to understand the role of the FAAH variant and the eCB system more broadly in neurodevelopment and psychiatric risk.